Dynamic Traffic Rearrangement to Enforce Policy Changes in Mpls Networks

1. A method of controlling traffic of data through a network, the network having a plurality of Label Switched Paths (LSPs) created with Multi-Protocol Label Switching (MPLS), the method comprising: (a) monitoring network traffic through the plurality of LSPs and determining utilization of LSPs on a DiffServ class basis; (b) if LSP utilization for one or more classes exceeds a predetermined threshold, issuing a rearrange request comprising: (i) a new policy on network bandwidth fraction to be allocated for each DiffServ class (c) computing LSP rearrangement in response to a rearrange request wherein a list of LSPs reconfigured includes one or more LSPs from which traffic is to be rerouted to other LSPs.

2. The method of claim 1 wherein the list of LSPs reconfigured includes a list of data flows which are reassigned to other LSPs.

3. The method of claim 1 wherein the least amount of LSPs is created in order to accommodate the rearrange request.

4. A method of controlling traffic of data through a network, the network having a plurality of Label Switched Paths (LSPs) created with Multi-Protocol Label Switching (MPLS), the method comprising: (a) monitoring network traffic through the plurality of LSPs and determining utilization of LSPs on a DiffServ class basis; (b) if LSP utilization for one or more classes exceeds a predetermined threshold, issuing a rearrange request comprising: (i) a new policy on network bandwidth fraction to be allocated for each DiffServ class, wherein the rearrange request is issued to a programming module which (i) reconfigures some LSPs by reducing their provisioned and/or allocated bandwidth, (ii) creates one or more LSPs, and (iii) reassigns data flows to the one or more new LSPs.

5. A method of controlling traffic of data through a network, the network having a plurality of Label Switched Paths (LSPs) created with Multi-Protocol Label Switching (MPLS), the method comprising: (a) monitoring network traffic through the plurality of LSPs and determining utilization of LSPs on a DiffServ class basis; (b) if LSP utilization for one or more classes exceeds a predetermined threshold, issuing a rearrange request comprising: (i) a new policy on network bandwidth fraction to be allocated for each DiffServ class (c) if LSP utilization for one or more DiffServ classes falls below a predetermined threshold, issuing a rearrange request comprising: (i) a new policy on network bandwidth fraction to be allocated for each DiffServ class, wherein the rearrange request is issued to a programming module which (ii) deletes one or more LSPs, and (iii) reassigns data flows previously assigned to the deleted LSPs.

6. A method of traffic engineering comprising: (a) classifying network traffic into different classes; (b) assigning a data forwarding policy to each of the classes; (c) assigning a plurality of paths to each of the classes; (d) periodically monitoring flow of traffic through the paths; (e) if the flow of traffic through any of the paths exceeds a first predetermined limit, reassigning certain traffic to different paths; and (f) if the flow of traffic through any of the paths falls below a second predetermined limit, destroying the paths with traffic below the second predetermined limit and reassigning the traffic assigned to the destroyed paths.

7. The method of claim 6 wherein the paths are label switched paths.

8. The method of claim 7 wherein traffic with the highest preference are never destroyed to the extent possible.

9. The method of claim 6 wherein the step of assigning a plurality of paths to each of the classes further comprises assigning a number of paths to a class based on the class's preference such that the class with the highest preference is assigned the most paths.

10. The method of claim 6 wherein the step of reassigning traffic to new paths if the flow of traffic through any of the paths exceeds a first predetermined limit further comprises creating at least one new path.

11. The method of claim 6 wherein the step of reassigning traffic to new paths if the flow of traffic through any of the paths exceeds a first predetermined limit further comprises creating at least one new path if traffic through a predetermined number of paths exceeds the first predetermined limit.

12. The method of claim 6 wherein the step of reassigning traffic to different paths if the flow of traffic through any of the paths exceeds a first predetermined limit comprises reassigning data traffic currently assigned to the path or paths that exceeds the first predetermined limit previously to paths that are assigned to classes with a lesser preference.

13. A system of traffic engineering comprising: (a) means for classifying network traffic into different classes; (b) means for assigning a data forwarding policy to each of the classes; (c) means for assigning a plurality of paths to each of the classes; and (d) means for periodically monitoring flow of traffic through the paths wherein if the flow of traffic through any of the paths exceeds a first predetermined limit, reassigning certain traffic to different paths; and if the flow of traffic through any of the paths falls below a second predetermined limit, destroying the paths with traffic below the second predetermined limit and reassigning the traffic assigned to the destroyed paths.

14. The system of claim 13 wherein the paths are label switched paths.

15. The system of claim 14 wherein paths with traffic with the highest preference are never destroyed to the extent possible.

16. The system of claim 13 wherein said means for assigning a plurality of paths to each of the classes further comprises means for assigning a number of paths to a class based on the class's preference such that the class with the highest preference is assigned the most paths.

17. The system of claim 13 wherein reassigning traffic to different paths if the flow of traffic through any of the paths exceeds a first predetermined limit further comprises creating at least one new path.